For knowledge to be managed it has to be severed from those who produced it; it must be stable, replicable, and translatable across contexts, space and time. What this entails is that at some point in its development it has to be divided from its auspices as a specific knowledge of specific people. In science the norms of replication and experimentation enable this division. In the commercial world, where what is required is a commercial product that can be marketed as distinct, different norms operate. In this paper we explore what we take to be a significant way of making such division, which entails the strategy of differentiating that which is ‘soft’ from that which is ‘hard’. Such categories are not self evident and are always socially constructed. In this paper we look at the process through which the division is made up.